Influenza Virus Surveillance in Coordinated Swine Production Systems, United States

To clarify the epidemiology of influenza A viruses in coordinated swine production systems to which no animals from outside the system are introduced, we conducted virologic surveillance during September 2012–September 2013. Animal age, geographic location, and farm type were found to affect the prevalence of these viruses.

I nfluenza A viruses (IAVs) are the etiologic agents of acute respiratory disease in many mammalian species. Although originating in wild aquatic birds, IAVs have been successful in crossing the species barrier, and specific subtypes have become endemic among humans and domestic swine populations (1). In the United States, influenza was first described in swine herds during the 1918 pandemic and has circulated among domestic pigs for nearly a century (2). The ability of swine IAVs to infect humans and cause pandemics such as that of the influenza A(H1N1) virus observed during 2009 (3,4) and the sporadic transmission of various swine influenza viruses, including H1N1 (5), H3N2 (6), and variant H3N2 (7), are public health concerns and highlight the need for increased vigilance and understanding of IAV epidemiology among swine.
Here we report the results from 13-months of active surveillance of IAV in coordinated swine production systems in the United States. The objectives of this study were 1) to determine the prevalence of IAV within farms in a closed production system and 2) to determine which sampled population is most affected by IAV.

The Study
Multisite coordinated production systems are the common method of swine production in the United States. These systems consist of multiple farms operating in tandem, with each farm responsible for 1 stage of the production process ( Figure 1). These systems are closed, meaning there are no introductions of animals from outside the system. Each farm in the system has a specific purpose: 1) to breed, gestate, farrow, and raise to the point of weaning replacement breeding stock of a specified genotype (multiplier farms [MF]); 2) to raise replacement female pigs, commonly called gilts, to 5-6 months of age for breeding (gilt development farms or units [GDU]); and 3) to breed, gestate, deliver, and raise to the point of weaning piglets specifically for meat production (breed-to-wean farms [BTW]). Gilts from GDUs are moved to MF or BTW farms for breeding. All farms house multiple age cohorts, although in different rooms or buildings. During suckling of piglets on MF and BTW farms, as well as in GDU farms, 1 cohort of piglets differing in age by <1 week are housed in a single room/building, and 1 cohort is removed before the entry of the next.
Four coordinated, multisite production systems, each consisting of 1 MF farm that includes both gilts and pigs to produce replacement female stock, 1 GDU farm to raise replacement female gilts from 3 to ≈26 weeks of age, and 4 BTW farms to raise pigs for meat, were selected to monitor the dynamics of IAV transmission in swine breeding herds. Systems are located across the United States. System 1's MF and GDU sites are located in Illinois and its BTW sites in Georgia. Systems 2, 3, and 4 are located entirely in Illinois, Oklahoma, and Nebraska, respectively.
Nasal swab samples were collected from pigs monthly during September 2012-September 2013 from each farm in all 4 systems. The animal type and time of sampling differed on the different farm types. On the MF and BTW farms, samples were collected from 30 piglets at ≈3 weeks of age and from 30 gilts. On the MFs, 30 gilts were sampled before entry to the farm. On the GDU farms, 30 animals at ≈26 weeks of age were sampled before movement to a BTW. In the BTWs, gilts were sampled before breeding (4-8 weeks after arrival) and piglets were sampled immediately before weaning ( Figure 1). This strategy enabled assessment of IAV status before the movement of animals to the next stage of production. Laboratory methods are summarized in the online Technical Appendix (http://wwwnc. cdc.gov/EID/article/21/10/14-0633-Techapp1.pdf).
During the 13-month period spanning September 2012-September 2013, a total of 14,954 swab samples were collected and tested for the presence of the IAV matrix gene by real-time reverse transcription PCR. Of the samples collected, 741 (5.0%) tested positive, which is consistent with previous surveillance studies (8).
Bivariate analysis found statistically significant correlations between infection and location in Illinois, GDU  Swine Influenza in Coordinated Production Systems farm type, and system 2 (p<0.001 for all), but not for age (Table 1). We then constructed a logistic regression model that assessed the effect of age, system, location, and farm type on having a positive influenza result. Age was statistically significant by this model (p = 0.004); the odds ratio for piglets at weaning whose samples tested positive for IAV was 1.3 (95% CI 1.1-1.6) compared to that for gilts (Table  2), which is consistent with previous studies (9). However, this finding could be related to interaction between age and state/system. System 2 again had higher odds of positive  (Figure 2). All 3 common porcine influenza subtypes (H1N1, H1N2, and H3N2) were detected Figure 1. Flow of animals through a closed swine production system. Coordinated swine production systems maximize desired animal traits and weight gain. High-quality breeding sows grown and bred in multiplier farms (MF-Gilts) were sampled transfer to commercial gilt development farms (GDU), where they were sampled again at exit (6 months). At ≈3 weeks of age, piglets were sampled before weaning (MF-Pigs). Mature gilts were transported from the GDU to 1 of 4 commercial breed-to-wean (BTW) farms, where samples were collected from gilts before entry and piglets before weaning. during the 13-month surveillance period. Increased prevalence of IAV was detected in piglets (MF pigs, GDU, BTW pigs) in all 4 systems, particularly in systems 1, 2, and 4, from winter to early summer ( Figure 2, panels A, B, D), which is consistent with other studies (8). Multiple subtype detection occurred only on BTW farms in system 2 ( Figure  2, panel B).

Conclusions
We found that IAV infection was present at all stages of swine production within coordinated production systems. Animal age, geographic location, and type of farm affected risk for infection. We also found continuing virus circulation in all populations year round, although prevalence was higher from winter through early summer. To fully elucidate the factors that contribute to persistent IAV infection in swine farms and therefore develop evidence-based control strategies, further research is needed.